Various transactions and web services may be accomplished using computing platforms based on virtual machines. As a result, in the event of a virtual machine (“VM”) failure, the time to recover the VM may be a business-critical performance factor. The recovery of a VM from a backup archive may be time consuming, especially if there is a large amount of data to be transferred from the backup archive to its recovery destination. As such, the services provided by the VM may be unavailable while the VM is being recovered after failure. Accordingly, reducing the time between the failure of a VM and its recovery may be a primary goal in various business contexts.